Pump Club
Pump Club

Web App

A social exercise web app that uses Tensorflow to keep count and form for bodyweight exercise.
Designer
Developer
Figma
Blender
React
Javascript
HTML
CSS

According to a 2020 CDC study, only 24.2% of Americans meet guidelines for both aerobic and muscle-strengthening activities. There is a downward trend in physical fitness along with increases in mental health issues.

I set out to develop a prototype app for individuals and their communities to build and maintain powerful calisthenics habits in manageable and accessible chunks, wherever they are and however busy they may be.

Research

Being a strong fitness and strength training advocate, push up and squat tracking methods seemed key especially for their accessibility and whole body benefits. You can do them anywhere and easily scale difficulty and benefits with difficulty of technique or number of reps. A beginner may do a few knee to ground push ups and may eventually advance to do 20 standard push ups with varying tempo while controlling tension.

First, we must seek existing methods that others have developed for tracking exercise and their 1. Technical feasability (Can I achieve the same alone? With others?) and 2. Accuracy (Can it count reliably? Can people cheat e.g. failsafe?).
I find several attempts at solving for push ups. One involves having a phone below your face and tapping your nose on the screen to count (What?). Another advances this with a proximity sensor using the camera. I love the UX of this but the problem is inaccurate tracking most likely by the user e.g. cheating. Vingroup's push-up counter made in OpenCV supersedes this all ways. It uses camera and AI, so there's endless conditions for accuracy, make it predictive, or give detailed data.

Design

This is where I showcase my designs and explain those decisions

Evaluation

Review of designs at this point and personal and peer thoughts